HAL Id: hal-00891855
https://hal.archives-ouvertes.fr/hal-00891855
Submitted on 1 Jan 2004
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of
sci-entific research documents, whether they are
pub-lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
cornuta (Hymenoptera: Megachilidae) and Apis
mellifera (Hymenoptera: Apidae) on ”Comice” pear
Víctor Monzón, Jordi Bosch, Javier Retana
To cite this version:
DOI: 10.1051/apido:2004055
Original article
Foraging behavior and pollinating effectiveness of Osmia
cornuta (Hymenoptera: Megachilidae) and Apis mellifera
(Hymenoptera: Apidae) on “Comice” pear
Víctor H. M
ONZÓNa,b, Jordi B
OSCHc,b*, Javier R
ETANAba Laboratorio de Inmunología y Genética Molecular, Instituto de Biología,
Universidad Católica de Valparaíso, Chile
b Unitat d’Ecologia, Facultat de Ciències, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain c Biology Department, Utah State University, Logan, UT 84322-5310, USA
(Received 3 April 2003; revised 20 January 2004; accepted 5 February 2004)
Abstract – We studied the pollinating effectiveness of Osmia cornuta and Apis mellifera on ‘Comice’ pear. Osmia cornuta visited more flowers per minute (13.8) than A. mellifera (7.1–9.8). Both species visited
similar numbers of flowers per tree (6.7–7.9), and switched rows with similar frequency (4.0–7.9%). Rate of stigma contact was 98.7% for O. cornuta, 51.8% for A. mellifera pollen-nectar foragers, and 19.0% for
A. mellifera nectar foragers. Fruit-set in flowers visited once was 28.9, 29.3, and 12.9%, respectively.
Seed-set was similar (7–8 seeds per fruit) in all three types of flowers. Osmia cornuta females provisioned their nests with 94.4% pear pollen. Stepwise multiple regression showed that fruitlet-set across the orchard was mostly related to tree size (trees with fewer flowers set more fruit per flower) (β= –0.64; P < 0.001), whereas seed-set at harvest was best explained by O. cornuta (β= 0.47; P < 0.005) and A. mellifera (β= 0.35; P < 0.05) spatial distribution.
Osmia cornuta / Apis mellifera / pollinating effectiveness / pear pollination
1. INTRODUCTION
Most commercial cultivars of pear, Pyrus
communis L., are self-incompatible or produce
more and/or better fruit when cross-pollinated (Nyéki, 1996). For this reason, rows of polli-nizer cultivars are normally planted among rows of the main cultivar in pear orchards, and 2.5–4 honey bee, Apis mellifera L., colonies per ha are introduced to ensure adequate pollen transfer (McGregor, 1976; Free, 1993; Benedek, 1996). However, insufficient pollination is a common cause of poor yields in pear (Sotes Ruiz, 1977; Soltész, 1996a). The sugar content of pear nectar is low (often < 10%) compared to that of other fruit tree species with
overlap-ping flowering periods (apple, Malus
domes-tica Borkh, peach, Prunus persica L., cherry, Prunus avium L.) (McGregor, 1976; Free,
1993; Soltész, 1996b). Apis mellifera workers visiting pear blossoms often switch to other fruit tree species or to other, more rewarding plants including Sinapis alba L., Stellaria spp.,
Taraxacum officinale Webber and Crataegus monogyna Jacq. when these start to bloom
(review in Free, 1993).
“Comice” pear is highly self-incompatible (Callan and Lombard, 1978), and typically has poor productivity, even when planted in orchards with cross-compatible cultivars (Joumien, 1968; Bini and Bellini, 1971; Callan and Lombard, 1978). The poor productivity of “Comice” pear
is at least partially explained by its brief effective pollination period, which lasts only 1–2 days at 9–10 °C (Lombard et al., 1971; Crisosto et al., 1988). In other words, although stigmas remain receptive for longer periods, pollination needs to be accomplished in the first 1–2 days after anthesis for fertilization to occur before ovule degeneration (Nyéki and Bubán, 1996). Parthe-nocarpy in “Comice” is not as common as in other pear cultivars (Stephen, 1958; Joumien, 1968).
Osmia cornuta (Latreille) is an early-flying
mason bee that has been studied as an orchard pollinator (Asensio, 1984; Torchio and Asensio, 1985; Krunic et al., 1991; Bosch, 1994a, b).
Osmia cornuta is strongly attracted to fruit tree
flowers, and nests produced in orchards contain 78–99% pollen of the target crop (Márquez et al., 1994; Vicens and Bosch, 2000a; Maccagnani et al., 2003). In a recent study in a pear orchard pollinated almost exclusively with O. cornuta, seed-set was comparable to seed-set in trees caged with abundant pollinators, and, in some cultivars/years, fruit-set was found to decrease at increasing distances from the O. cornuta nesting shelters (Maccagnani et al., 2003). The pollinating effectiveness of O. cornuta is near 100% on almond and apple (Bosch and Blas, 1994; Vicens and Bosch, 2000a). To our knowledge, the foraging behavior and individ-ual pollinating effectiveness of O. cornuta on pears have not been previously reported. The aim of our study was to determine whether O.
cornuta will establish in pear orchards and
pro-vision nests with pollen of this species, and to study its foraging behavior and pollinating effectiveness on pear in comparison to A.
mel-lifera. We were also interested in determining
whether it is possible to sustain O. cornuta pop-ulations in pear orchards.
2. MATERIALS AND METHODS 2.1. Study orchard and flowering
phenology
Field work was conducted in April 1996 at Can Oller, a 100-ha pear plantation in Cassà de la Selva (Girona, NE Spain; 46° 37’ N, 4° 90’ E). We worked at a 3.8-ha orchard with 28 rows of 300 trees each. The main cultivar (“Comice”) was planted in 3 blocks of 8 rows each. “Comice” blocks were sepa-rated from each other by 2 rows of pollenizers (“Duc de Bordeaux” and “Conference”; Fig. 1). Distance
between rows was 3.75 m, and distance between adjacent trees within a row was 1.2 m, so that their limbs sometimes overlapped. Trees were eight years old, approximately 2.5 m tall and pruned in a central leader system. Adjacent orchards were planted with the cultivar “Passa Crassana”. The flowering phe-nology of the cultivars “Comice”, “Duc de Bordeaux”, “Conference” and “Passa Crassana” was estimated on 4 tagged limbs (oriented to N, S, E and W, respec-tively) on each of 10 trees of each cultivar. Every morning throughout the flowering period, we counted the number of new open flowers on each tagged limb. Honey bee colonies were introduced across the Can Oller farm at a rate of 5 colonies (with 4 frames of brood) per ha. A group of 8 colonies was located 20 m away from the southeast corner of the study orchard.
Figure 1. “Comice” block showing pollenizer
2.2. Management of Osmia cornuta populations
Five O. cornuta nesting shelters were set up in five adjacent rows in the middle of the “Comice” orchard (Fig. 1). Each shelter consisted of a wooden box with the front side open held at 1.5 m above the ground by 4 metal fence posts. Each shelter con-tained 20 solid wood blocks with 25 drilled holes per block (Bosch, 1995). A paraffin-coated paper straw (15 cm long, 8 mm inside diameter) was inserted in each hole. On 3 April, an O. cornuta population con-sisting of 707 females and 1996 males was released in the orchard. Most bees (656 females and 1708 males) had been reared from trap-nests (wood blocks identical to those described above) placed in several locations near Girona in 1995. The remaining bees (51 females and 288 males) were the progeny of a population managed for almond pollination in Mur-cia (SE Spain). Both groups of bees had been reared at 25–30 °C during spring-summer of the previous year. By mid September, when bees had reached the adult stage, paper straws were dissected. Cocoons were then removed and sexed, female cocoons being larger and typically located in the innermost cells within a nest (Bosch, 1994b). Cocoons were then wintered at 3 °C until the following spring, when female cocoons were individually inserted in paper straws, which were evenly distributed among nest-ing wood blocks in the five nestnest-ing shelters (~140 females per shelter). Male cocoons were placed in open cardboard boxes set in the nesting shelters. Cocoons from the Murcia population were kept sep-arate from those of the Girona population. Every other day after release, straws with inserted female cocoons were inspected to determine female emer-gence. On 11 and 16 April, all nesting cavities were inspected at night with a flashlight to count the number of females established in the nesting wood blocks.
Osmia cornuta nesting materials were removed
from the nesting shelters in early May, after petal fall. Straws in which female cocoons had been inserted for release were examined to determine female emergence percentages. Paper straws con-taining O. cornuta progeny were incubated at 25 °C throughout the summer. At the end of August, when progeny had reached the adult stage, each nest was dissected and its contents analyzed. Cells were counted and sexed as described above for the paren-tal population. Developmenparen-tal morparen-tality and para-sitism were also recorded. Fecal samples from 110 randomly-chosen cells were mounted on fuchsin-stained gelatin microscope slides (Beattie, 1971), and 300 pollen grains per cell were identified to genus or tribe level under the microscope (100 X) using a reference pollen collection. Because pear was the only fruit tree in bloom in Can Oller through-out the study and most O. cornuta females forage
within 200 m from their nesting site (Vicens and Bosch, 2000a, b), all pollen identified as Rosaceous fruit tree was assumed to be pear.
2.3. Pollinator counts
Pollinator counts were made on 32 “Comice” trees (8 trees of each of 4 rows) located at increasing dis-tances from the O. cornuta nesting shelters (Fig. 1). All 4 rows sampled were 4 rows away from a pol-lenizer row. The first four trees sampled in each of the four rows were 12 m apart from each other (10 tree-intervals). The last four trees in each row were 24 m (20 trees) apart from each other. Counts were conducted on 10 days throughout the flowering period of “Comice”. Four counts were completed daily at 0900–1000, 1200–1300, 1500–1600, and 1800–1900 h local time. During each count, each tree was observed for 1 min, and all insects seen visiting flowers were recorded. To avoid altering visitation frequencies, recorded insects were not captured.
2.4. Foraging behavior and pollinating effectiveness
Data on foraging behavior were obtained at “Comice” peak bloom and on days with weather favorable to both O. cornuta and A. mellifera activ-ity. Female O. cornuta and worker A. mellifera were followed from their first flower visit to a tree until we lost sight of them. Data were gathered on: (1) resources collected (nectar and/or pollen); (2) stigma contact (pear blossoms have 5 stigmas and it is dif-ficult to ascertain how many are contacted, so we scored visits as either contacting some stigmas or none); (3) visitation rates (flowers visited per minute); (4) number of flowers visited per tree; and (5) rela-tive position of consecurela-tively visited trees (adjacent trees in the same row, other trees in the same row, trees in other rows).
Pollinating effectiveness was measured over four consecutive days (20–23 April) by letting non-vis-ited flowers of trees located in the same rows as those used to count pollinators, be visited once by either
O. cornuta or A. mellifera. “Comice” limbs with
flower buds were covered with fine mesh nylon bags (0.5 mm-openings) to prevent insect visitation and inspected daily. When open flowers were available (first day of anthesis), they were unbagged and watched continuously until either a female O.
cor-nuta or a worker A. mellifera visited a flower. Visited
On 20 April, shortly before bloom initiation, we counted all inflorescence buds in the same 32 “Comice” trees used for the pollinator counts. We then multiplied the number of inflorescence buds by 5 to obtain an estimate of number of flowers per tree. Fruitlets produced in these trees were counted on 20 May, and then again on 15 September, at harvest. At that time, a sample of 25 fruits per tree was used to count seeds per fruit. Thus, we obtained a measure of fruitlet-set (developing fruits retained on the tree before the June drop/flowers produced), and seed-set (seeds/fruit retained on the tree at harvest) across the orchard.
2.5. Data analysis
We used G-tests to analyze differences between pollinator species in stigma contact rate and fre-quency of flights between rows. Differences in vis-itation rates, flowers visited per tree and seed-set in single-visit flowers were analyzed with one-way ANOVA. Results obtained over the four days on which pollinating single-visit measurements were conducted were pooled together because there were no significant differences across dates in fruit-set, irrespective of whether all visits were considered (G = 4.631; df = 3; P > 0.2) or only visits with stigma contact were considered (G = 3.801; df = 3; P > 0.2). Seed number variances across pollinator type were homogeneous (Levene test: F = 0.249; df = 2.52;
P > 0.7). To evaluate the effect of insect visitation
on fruitlet-set and seed-set in the 32 trees sampled,
we used Stepwise Multiple Regression. The inde-pendent variables were number of flowers per tree and frequency of A. mellifera, O. cornuta and other flower-visiting insects, respectively. Number of flowers per tree was included because preliminary exploration of the data showed a significant negative correlation between this variable and fruitlet-set. Multiple regression provides an estimate of the rel-ative magnitudes of the contributions of the inde-pendent variables to the observed variation in the dependent variable (Sokal and Rohlf, 1995). Toler-ance of all four independent variables was > 0.58 in all analyses, indicating absence of excessive colline-arity (Quinn and Keough, 2002).
3. RESULTS
3.1. Osmia cornuta emergence
14 April. Pollen from pollenizer cultivars “Conference”, “Duc de Bordeaux” and “Passa Crasana“ was available throughout the 13-day flowering period of “Comice” (Fig. 1).
3.2. Pollinator counts
A total of 950 insects were recorded visiting “Comice” flowers. Apis mellifera was, by far, the most abundant species (55.4% of the insects recorded). Osmia cornuta and wild bees amounted to 7.9% and 2.4% of the visits recorded, respec-tively. Most (88%) O. cornuta recorded were female. Among the Diptera, the Syrphidae were particularly abundant (28.6%). Other Diptera and other insects (mainly aculeate wasps and Lep-idoptera) amounted to 5.3% and 0.4% of the visits recorded, respectively. Overall pollinator density was 0.74 insects per tree per minute.
3.3. Foraging behavior and pollinating effectiveness
Of the 110 O. cornuta fecal samples exam-ined, 75.5% contained exclusively pear pollen, and 23.6% contained more than 80% pear pol-len. Overall, 94.4% of the pollen grains identi-fied were pear. Other pollens found were Erica,
Quercus, Brassica, and Asteraceae
(Sonchus-type). Nesting O. cornuta females collect pol-len and nectar simultaneously on pear flowers. They nearly always land on the reproductive organs of the flower and insert their proboscis to probe for nectar while they scrape the anthers with their hind legs to collect pollen on their ventral scopa. Contact of the bee and the stig-mas was observed on 99% of the visits (Tab. I).
On most visits, all or most stigmas were con-tacted. Osmia cornuta males and pre-nesting females only collect nectar for their own con-sumption. However, males and pre-nesting females were also observed to contact the stig-mas of pear flowers on most visits. Of the A.
mellifera recorded, 40.7% collected only
nec-tar. These individuals typically landed on the petals of the flower and very often probed the nectaries without contacting the stigmas (Tab. I). Other A. mellifera (59.3%) collected both nectar and pollen. These individuals (henceforth, pollen foragers) also tended to land on the petals of the flower, but they made stigma contact on 52% of their visits, as they scraped the anthers to collect pollen (Tab. I). Differences in visit effectiveness between the three pollinator types were statistically signif-icant (G = 734.5; df = 2; P < 0.0001). Flower visitation rates also differed between pollinator types (F = 10.13; df = 2,104; P < 0.0001; Tab. I). Instead, all three pollinator types visited similar numbers of flowers per tree (F = 0.47; df = 2,120; P = 0.62; Tab. I). Both O. cornuta and
A. mellifera tended to fly from one tree to an
adjacent tree in the same row, and differences between pollinators in inter-row flight fre-quency were not significant (G = 0.50; df = 1;
P = 0.48; Tab. II).
Fruitlet-set of single-visited flowers was more than twice as high for flowers visited by
A. mellifera pollen-foragers or O. cornuta as
for flowers visited by A. mellifera nectar-for-agers (G = 7.62; df = 2; P < 0.03; Tab. III). The number of seeds per fruit did not differ signif-icantly between flowers visited by the three pollinator types (F = 1.13; df = 2,52; P = 0.33; Table I. Visit effectiveness (visits in which at least one stigma was contacted), visitation rates, and number
of flowers visited consecutively per tree in Osmia cornuta females and Apis mellifera workers on “Comice” pear.
Effective visits Visitation rates
Tab. III). Rate of stigma contact on the previ-ously-bagged flowers used in this experiment (Tab. III) was similar to that of open-pollinated flowers (Tab. I) for O. cornuta. Instead, stigma contact was higher in previously-bagged flowers than in open-pollinated flowers for A. mellifera, especially for pollen-foragers (Tabs. I and III). Although some O. cornuta were recorded as far as 120 trees (120 m) from the nesting shel-ters, most (78.7%) foraged within the first 40 trees (48 m) nearest the nesting shelters. In contrast, A. mellifera (47.5% recorded within the first 40 trees) and other pollinators (45.8% recorded within the first 40 trees) were more evenly distributed throughout the orchard (Fig. 3).Mean fruitlet-set in the 32 trees sam-pled was 23.6 ± 2.3%, fruit-set at harvest 4.6 ± 0.3%, and seed-set 2.9 ± 0.1 seeds per fruit. The only independent variable included in the Step-wise Multiple Regression model for fruitlet-set (r2 = 0.405; F = 20.39; df = 1,30; P < 0.0001)
was number of flowers per tree (β = –0.636; P < 0.0001). Thus, small trees (with fewer flowers) tended to have greater fruitlet-set, irrespective of visitation by different pollinator types. The model for seed-set (r2 = 0.362; F = 8.22; df =
2,29; P < 0.005) included O. cornuta frequency (β = 0.47; P < 0.005) and A. mellifera frequency (β = 0.35; P < 0.05).
3.4. Osmia cornuta population dynamics
Osmia cornuta female emergence was high
(94.2%) for the Girona population, but low (21.6%) for the Murcia population. Overall, 629 females emerged, of which 282 (44.8%) were recorded in the night counts made one week after release. By the end of the flowering period, these females had produced 496 nests, of which 24 were incomplete (unplugged). The total number of cells produced was 2138, of which 675 contained female and 1463 male progeny (sex ratio = 2.17 / ). Therefore, an average of 2.39 female cells, and 5.19 male cells were produced per nesting female. Mor-tality was similar for progeny of either sex (14.3% of the cells produced). Most mortality occurred in the early-immature (egg, first and second larval instars) and prepupal stages (7.4% and 4.7%, respectively). Parasitism, was rare (0.28%), and due to a single cleptoparasitic species, the drosophilid fly Cacoxenus
indaga-tor Loew.
4. DISCUSSION
As with other rosaceous fruit trees, A. mellifera pollen-foragers were much more effective pol-linators of “Comice” pear than nectar-foragers Table II. Flights between pear trees by Osmia cornuta females and Apis mellifera workers (both nectar and
pollen-nectar foragers). Species n Adjacent trees in same row Other trees in same row Trees in other rows n % n % n % O. cornuta 89 78 87.6 4 4.5 7 7.9 A. mellifera 25 22 91.7 2 8.0 1 4.0
Table III. Fruit- and seed-set in “Comice” flowers receiving a single visit by Osmia cornuta females or Apis mellifera workers.
(Robinson, 1979; Bosch and Blas, 1994; Vicens and Bosch, 2000a). Stigma contact in previ-ously unvisited flowers was similar to that in open-pollinated flowers for O. cornuta. In trast, A. mellifera were much more likely to con-tact the stigmas on previously unvisited flowers than on open-pollinated flowers. This result shows that single-visit experiments may some-times overestimate pollinating effectiveness. Because they have not been visited, virgin bagged flowers are expected to have more pollen and nectar than the average open-pollinated flower. For Diervilla lonicera, accumulation of nectar in virgin bagged flowers resulted in prolonged flower-handling times and increased pollen dep-osition on the stigma (Thomson and Plowright, 1980). In addition, previously bagged flowers may experience a different microclimate com-pared to unbagged flowers, which may influ-ence pollen and nectar presentation (Kearns and Inouye, 1993). The fact that seed-set in pre-viously bagged flowers was high (7–8 seeds) and similar for all three pollinator types indi-cates that both O. cornuta and A. mellifera car-ried sufficient amounts of viable, compatible pollen for adequate fertilization, as long as stigma contact occurred. It also suggests that seed-set in these flowers may have been limited by phys-iological constraints or nutritional resources,
rather than pollen deposition. Similar results have been obtained on almond and apple (Bosch and Blas 1994; Vicens and Bosch, 2000a). Taken together, these results question the use-fulness of pollen dispensers placed at the hive entrance for orchard pollination with A.
mellif-era, and emphasizes the need to increase the
proportion of pollen-foragers (with higher stigma contact rates).
Commercial size in “Comice” pears is achieved in fruits with 2–3 seeds (Joumien, 1968; Callan and Lombard, 1978). Mean seed-set was 2.9 in fruits produced throughout the orchard from flowers exposed to open pollina-tion, in contrast to 7–8 in fruits produced in the single-visit experiment. This difference suggests again that pollinator effectiveness may be over-estimated in single-visit experiments due to longer visitation times in previously unvisited flowers, which would result in greater pollen deposition (Thomson and Plowright, 1980). In addition, some open-pollinated flowers could have been visited only by dipterans or other less effective pollinators.The difference in seed-set between open-pollinated fruits and single-visit fruits was not due to parthenocarpy in open-pollinated flowers, since we found no fruits with 0 seeds.
Figure 3. Distribution of Osmia cornuta, Apis mellifera and other flower-visiting insects throughout the
Osmia cornuta emergence was quick and
well timed with pear bloom. Female O. cornuta established in the nesting materials provided, and, in agreement with results of Maccagnani et al. (2003), almost exclusively visited pear flowers. Just as with almond and apple (Bosch and Blas, 1994; Vicens and Bosch, 2000a), stigma contact rate and pollinating effective-ness of O. cornuta females on pear were very high. Despite the small size of the population released, the contribution of O. cornuta to “Comice” pollination in the study orchard was notable. Multiple Regression analysis showed that a significant fraction of seed-set variation in “Comice” could be explained by the distri-bution of O. cornuta in the orchard, even when honey bees were seven times more abundant. From a commercial stand point, seed-set is an important parameter, since it is positively corre-lated with fruit size and regular shape (Joumien, 1968; Callan and Lombard, 1978). Together with the study of Maccagani et al. (2003), our results strongly indicate that O. cornuta should be considered a valid alternative pollinator of pears. Despite their abundance, syrphid flies contributed little to overall pollination.
Osmia cornuta stocking rates for adequate
apple pollination have been estimated at 355– 530 nesting females per ha (Vicens and Bosch, 2000a). This estimate was obtained assuming an average flower production of ~2000 flowers per tree and an average tree density of ~2000 trees per ha. In our “Comice” orchard, flower production was 1965 ± 104 flowers per tree and tree density was 2222 trees per ha. Osmia
cor-nuta rate of stigma contact, flower visiting rate,
fruit-set after one visit, and visitation fidelity (based on the pollen composition found in the provisions) on “Comice” were similar to values obtained on apple (Vicens and Bosch, 2000a). Thus, recommended stocking rates for pear pollination with O. cornuta would not be very different from those estimated for apple. How-ever, because pears bloom ahead of apples (with greater risk of unfavorable weather) and because blooming periods tend to be shorter for pear than apple, slightly higher numbers of nesting females may actually be required for pear pollination. Recommended stocking rates for A. mellifera on pear are 2.5–4 colonies per ha (McGregor, 1976; Free, 1993; Benedek, 1996). Recent studies have discussed the pos-sibility that a diverse pollinator assemblage
contributes to overall pollination in more ways than just by increasing pollinator abundance (Klein et al., 2003). Different pollinator species might complement each other in terms of spa-tial or temporal distribution. Or the foraging behavior of a pollinator may influence the pol-linating effectiveness of another pollinator. For example, yields in clover fields were some-times higher when nectar-robbing bumblebees were present (Heinrich, 1979). Presumably, lower nectar levels in fields with nectar robbers caused legitimate pollinators to move between plants more often, thus enhancing cross-polli-nation. We did not observe any specific inter-actions between O. cornuta and A. mellifera. Both species can be used, together or sepa-rately, to improve pear pollination. However, if used in conjunction, stocking rates will likely have to be adjusted. Overpollination will not result in higher yields, and floral resources may be lowered to levels that hinder Osmia larval development (Torchio, 1985) or enhance dis-persal of pre-nesting females (Bosch and Kemp, 2002).
The female O. cornuta population recovered at the end of the study (578 females) was 18.2% lower than the female population initially released, and 8.1% lower than the emerged female population. Several factors contributed to this population decrease. First, emergence of the Murcia population was very low (22%). Inspection of the cocoons with unemerged indi-viduals showed that most mortality occurred in the adult stage, and some cocoons were par-tially chewed, indicating that bees were alive when released, but not vigorous enough to come out of the cocoon. Similar observations were made in O. cornuta populations wintered too late in the autumn (that is, pre-wintered for too long) (Bosch, 1994a, 1995). In both O.
cor-nuta and the closely-related Osmia lignaria
dispersal is typical of populations released as single cocoons (Bosch, 1994c; Vicens and Bosch, 2000a). Better establishment (≥70%) can be obtained when populations are released from natal nests (Maeta, 1978; Torchio, 1984, 1985; Bosch, 1994c), and nesting shelters are distributed evenly across the orchard, rather than grouped in a small area (Torchio, 1984; Bosch, 1994a). Pears have a short blooming period compared to other fruit trees. More stud-ies will be needed to establish whether O.
cor-nuta populations can be routinely increased in
pear orchards, and if planting additional pollen-nectar sources within the orchard could help increase cell production by increasing the length of the nesting period (Bosch and Kemp, 2002).
ACKNOWLEDGEMENTS
We are grateful to P. Vilardell and J. Bonany (IRTA, Mas Badia, La Tallada) for their support and advice throughout the study, and to N. Vicens, M. Sala and J. Batchillí for invaluable assistance. M. Suárez-Cervera (Universitat de Barcelona) kindly determined the pollen samples, and A. Lacasa (CIDA, Murcia) provided some O. cornuta nests. We also thank M. Blas and A. Serra (Universitat de Barce-lona) for providing laboratory space and access to nesting materials. The manuscript was improved through comments from J.H. Cane (USDA/ARS, Logan, Utah), B.E. Vaissière (INRA, Avignon) and two anonymous reviewers. This study was sup-ported by a fellowship from the Instituto de Coop-eración Iberoamericana to V.H. Monzón.
Résumé – Comportement de butinage et efficacité pollinisatrice d’Osmia cornuta (Hymenoptera : Megachilidae) et d’Apis mellifera (Hymenoptera : Apidae) sur la poire « Comice ». La poire “Comice”
est un cultivar considéré comme difficile à polliniser en raison de son auto-incompatibilité et de la briè-veté de la période de pollinisation. Nous avons suivi dans un verger (Fig. 1) les butineuses d’Osmia
cor-nuta (cat. 1), les butineuses de pollen et de nectar
(cat. 2) et les butineuses de nectar seulement (cat. 3) d’Apis mellifera et enregistré les taux de contact avec le stigmate, les taux de visite (nombre de fleurs visi-tées par min), le nombre de fleurs visivisi-tées par arbre et les distances de vol entre les arbres. Les femelles d’O. cornuta ont émergé rapidement (Fig. 2). Elles ont visité plus de fleurs par min (13,8) qu’A.
melli-fera (7,1 à 9,8), mais les deux espèces ont visité le
même nombre de fleurs par arbre (6,7 à 7,9) et les fréquences de vols inter-rangs ont été semblables (4,0 à 7,9 %) (Tabs. I et II). Le contact avec le stig-mate a eu lieu dans 98,7 % des cas pour les
butineu-ses de catégorie 1 (O. cornuta), dans 51,8 % des cas pour la catégorie 2 et dans 19,0 % pour la catégorie 3 (Tab. I). Nous avons recouvert les boutons floraux de sacs en nylon puis ôté ces sacs pour permettre aux fleurs vierges d’être visitées une fois soit par O.
cor-nuta, soit par A. mellifera. La mise à fruits de ces
fleurs visitées une seule fois a été de 28,8 % pour les butineuses de catégorie 1, de 29,3 % pour la catégo-rie 2 et de 12,9 % pour la catégocatégo-rie 3 (Tab. III). Le contact avec le stigmate était plus élevé pour A.
mel-lifera chez les fleurs qui avaient été ensachées, alors
qu’O. cornuta se comportait de la même façon sur les deux types de fleurs (Tabs. I et II). La nouaison chez les fleurs visitées une seule fois et qui ont donné des fruits a été semblable pour les trois catégories de butineuses (7,3 à 8,2) (Tab. III). La plupart des femelles d’O. cornuta (78,7 %) ont butiné dans un rayon de 48 m autour de leur site de nidification (Fig. 3). Les nids d’O. cornuta ont été approvision-nés à 94,4 % en pollen de poirier. L’analyse de régression multiple pas à pas a montré que la mise à fruits dans le verger était principalement liée à la taille des arbres (les arbres avec moins de fleurs avaient une plus grande mise à fruits) (β = –0,64 ;
P < 0,001), alors que la nouaison s’expliquait mieux
par la répartition d’O. cornuta (β = 0,47 ; P < 0,005) et d’A. mellifera (β = 0,35 ; P < 0,05). On peut donc détecter la contribution d’O. cornuta à la pollinisa-tion bien que les butineuses d’ A. mellifera soient sept fois plus abondantes. La mortalité hivernale a été basse chez l’une des populations d’ O. cornuta (5,8 %) mais plus forte chez l’autre (78,4 %). Dans l’ensemble les femelles se sont installées lentement (44,8 % des femelles ayant émergé). Le résultat est que la population de femelles d’O. cornuta retrouvée à la fin de l’étude a été de 18,2 % inférieure à la popu-lation lâchée. La discussion porte sur une gestion pré-hivernale correcte afin de réduire la mortalité hivernale et sur des méthodes pour accroître l’ins-tallation des populations d’O. cornuta.
Osmia cornuta / Apis mellifera / efficacité
pollini-satrice / pollinisation / poirier
Zusammenfassung – Sammelverhalten und Be-stäubungseffizienz von Osmia cornuta (Hymenop-tera: Megachilidae) und Apis mellifera (Hyme-noptera: Apidae) an „Comice“ Birnen. Wir
untersuchten das Sammelverhalten und die Be-stäubungseffizienz von Osmia cornuta (Latreille) und Apis mellifera L. in einer “Comice” Birnenplan-tage (Abb. 1). Dieses Kultivat wird wegen seiner Selbstinkompatibilität und den kurzen effektiven Bestäubungzeiten als schwierig zu bestäuben ange-sehen. Wir folgten Sammlerinnen in der Birnenplan-tage und registrierten die Stigmakontaktraten, die Blütenbesuchsraten (Anzahl besuchter Blüten pro Minute), die Anzahl von aufeinanderfolgend besuchten Blüten pro Baum und die Flugstrecken zwischen besuchten Bäumen. Die Weibchen von
besuchten mehr Blüten pro Minute (13,8) als A.
mel-lifera (7,1–9,8), aber beide Bestäuberarten
besuch-ten ähnliche Anzahlen von Blübesuch-ten pro Baum (6,7– 7,9) und zeigten ähnliche Flugfrequenzen zwischen den Reihen, (4,0–7,9 %) (Tab. I und II). Der Stig-menkontakt war 98,7 % bei O. cornuta, 51,8 % bei
A. mellifera Pollen-Nektarsammlerinnen und 19,0 %
bei A. mellifera Nektarsammlerinnen (Tab. III). Wir bedeckten Blütenstände mit Nylonsäckchen und entfernten diese dann, um jungfräuliche Blüten einem einmaligen Besuch von O. cornuta oder von
A. mellifera auszusetzen. Der Fruchtansatz an diesen
einmal besuchten Blüten war 28,8 % bei O. cornuta, 29,3 % bei A. mellifera Pollen-Nektarsammlerinnen und 12,9 % bei A. mellifera Nektarsammlerinnen (Tab. III). Der Stigmenkontakt von A. mellifera war bei den zuvor eingesackten Blüten höher als bei offenbestäubten Blüten, während O. cornuta sich auf beiden Blütentypen ähnlich verhielt (Tab. I und III). Der Fruchtansatz bei einmal besuchten Blüten, der sich danach zu Früchten weiterentwickelte war bei allen drei Bestäubertypen ähnlich (7,3–8,2) (Tab. III). Die meisten (78,7 %) O. cornuta Weib-chen sammelten innerhalb von 48 m Entfernung von ihrem Nest (Abb. 3). Osmia cornuta Nester wurden zu 94,4 % mit Birnenpollen versorgt. Eine schritt-weise multiple Regressionsanalyse zeigte, dass der Fruchtansatz über die Plantage am meisten mit der Baumgröße zusammenhing (Bäume mit weniger Blüten hatte einen höheren Fruchtansatz (β = –0,64;
P < 0,001), während der Samenansatz am besten
durch die Verteilung von O. cornuta (β = 0,47; P < 0,005) und A. mellifera (β = 0,35; P < 0,05) erklärt wurde. Daher konnte der Bestäuberbeitrag von O.
cornuta noch erkannt werden, obwohl A. mellifera
siebenfach häufiger vertreten war. Die Wintersterb-lichkeit war in einer der eingesetzten O. cornuta Populationen niedrig (5,8 %), in der anderen aber hoch (78,4 %). Die Gesamterzeugung an Weibchen war niedrig (44,8 % der geschlüpften Weibchen). Als Resultat war die am Ende der Untersuchung wie-dergefundene Population von weiblichen O. cornuta um 18,2 % kleiner als die ausgesetzte Population. Wir diskutieren daher angemessenes Vorwinterma-nagement, um die Wintermortalität zu verringern und Methoden, um die Etablierung einer Population zu verbessern.
Osmia cornuta / Apis mellifera /
Bestäubungseffi-zienz / Birnenbestäubung
REFERENCES
Asensio E. (1984) Osmia (Osmia) cornuta Latr pol-linisateur potentiel des arbres fruitiers en Espagne (Hymenoptera, Megachilidae), in: Proc. V Symp. Int. Pollinisation, 27–30 September 1983, Ver-sailles, pp. 461–465.
Beattie A.J. (1971) A Technique for the study of insect-borne pollen, Pan. Pac. Entomol. 47, 82.
Benedek P. (1996) Insect pollination of fruit crops, in: Nyéki J., Soltész M. (Eds.), Floral biology of tem-perate zone fruit trees and small fruits, Akadémiai Kiadó, Budapest, pp. 287–340.
Bini G., Bellini E. (1971) Macrosporongenesi nella cultivar di pero “Decana di Comizio” e ricerca dil periodo più opportuno per la sua impollinazione, Riv. Ortiflorofrut. Ital. 4, 322–337.
Bosch J. (1994a) Improvement of field management of
Osmia cornuta (Latreille) (Hymenoptera,
Meg-achilidae), Apidologie 25, 71–83.
Bosch J. (1994b) The nesting behaviour of the mason bee Osmia cornuta (Latr) with special reference to its pollinating potential (Hymenoptera, Megachi-lidae), Apidologie 25, 84–93.
Bosch J. (1994c) Osmia cornuta Latr. (Hym., Meg-achilidae) as a potential pollinator in almond orchards, Releasing methods and nest-hole length, J. Appl. Entomol. 117, 151–157.
Bosch J. (1995) Comparison of nesting materials for the orchard pollinator Osmia cornuta (Hymenop-tera: Megachilidae), Entomol. Gen. 19, 285–289. Bosch J., Blas M. (1994) Foraging behaviour and pol-linating efficiency of Osmia cornuta and Apis
mel-lifera on almond (Hymenoptera: Megachilidae,
Apidae), Appl. Entomol. Zool. 29, 1–9. Bosch J., Kemp W.P. (2002) Developing and
estab-lishing bee species as crop pollinators: the exam-ple of Osmia spp. (Hymenoptera: Megachilidae) and fruit trees, Bull. Entomol. Res. 92, 3–16. Bosch J., Kemp W.P. (2004) Effect of pre-wintering
and wintering temperature regimes on weight loss, survival, and emergence time in the mason bee
Osmia cornuta (Hymenoptera: Megachilidae),
Apidologie 35, 469–479.
Bosch J., Kemp W.P., Peterson S.S. (2000) Manage-ment of Osmia lignaria (Hymenoptera: Megachi-lidae) populations for almond pollination: Meth-ods to advance bee emergence, Environ. Entomol. 29, 874–883.
Callan N., Lombard P. (1978) Pollination effects on fruit and seed development in “Comice” pear, J. Am. Soc. Hortic. Sci. 103, 496–500.
Crisosto C., Lombard P., Sugar D., Polito V. (1988) Putrescine influences ovule senescence, fertiliza-tion time, and fruit set in “Comice” pear, J. Am. Soc. Hortic. Sci. 113, 708–712.
Free J.B. (1993) Insect pollination of crops,Academic Press, London.
Heinrich B. (1979) Bumblebee economics, Harvard University Press, Cambridge.
Joumien F. (1968) The causes of poor bearing of pear trees of variety “Doyenne du Comice”, Acta Agro-bot. 21, 75–106.
the diversity of pollinating bees, Proc. R. Soc. London B 270, 955–961.
Kearns C.A., Inouye D.W. (1993) Techniques for pol-lination biologists, University Press of Colorado, Niwot.
Krunic M.D., Brajkovic M.M., Mihajlovic L.S. (1991) Management and utilization of Osmia cornuta Latr. for orchard pollination in Yugoslavia, Acta Hortic. 288, 190–193.
Lombard P., Cordy C., Hansen E. (1971) Relation of postbloom temperatures to “Bartlett” pear matu-ration, J. Am. Soc. Hortic. Sci. 96, 799–801. Maeta Y. (1978) Comparative studies on the biology
of the bees of the genus Osmia of Japan, with spe-cial reference to their management for pollination of crops (Hymenoptera: Megachilidae), Tohoku Natl. Agric. Exp. Stn. Bull. 57, 1–221 [in Japa-nese].
Maccagnani B., Ladurner E., Santi F, Burgio G. (2003)
Osmia cornuta (Latreille) (Hymenoptera:
Meg-achlidae) as a pollinator of pear (Pyrus communis L.): fruit- and seed-set, Apidologie 34, 207–216. Márquez J., Bosch J., Vicens N. (1994) Pollens col-lected by wild and managed populations of the potential orchard pollinator Osmia cornuta Latreille, J. Appl. Entomol. 117, 353–359. McGregor S.E. (1976) Insect pollination of cultivated
plants, Agricultural Handbook No. 496, ARS-USDA, Washington DC.
Nyéki J. (1996) Fertilization conditions, in: Nyéki J., Soltész M. (Eds.), Floral biology of temperate zone fruit trees and small fruits, Akadémiai Kiadó, Budapest, pp. 185–256.
Nyéki J., Bubán T. (1996) Pollination and fertilization, in: Nyéki J., Soltész M. (Eds.), Floral biology of temperate zone fruit trees and small fruits, Akadémiai Kiadó, Budapest, pp. 153–184. Quinn G.P., Keough M.J. (2002) Experimental design
and data analysis for biologists, Cambridge Uni-versity Press, Cambridge.
Robinson W.S. (1979) Effect of apple cultivar on for-aging behaviour and pollen transfer by honey bees, J. Am. Soc. Hortic. Sci. 104, 596–598.
Sokal R.R., Rohlf F.J. (1995) Biometry, W.H. Freeman and Company, New York.
Soltész M. (1996a) Requirements for successful fruit set in orchards, in: Nyéki J., Soltész M. (Eds.), Flo-ral biology of temperate zone fruit trees and small fruits, Akadémiai Kiadó, Budapest, pp. 257–286. Soltész M. (1996b) Flowering, in: Nyéki J., Soltész M. (Eds.), Floral biology of temperate zone fruit trees and small fruits, Akadémiai Kiadó, Budapest, pp. 80–131.
Sotes Ruiz V. (1977) Study on the pollination and fruit-set on “Blanca de Aranjuez” pear, Acta Hortic. 69, 235–241.
Stephen W.P. (1958) Pear pollination studies in Ore-gon, Bull. Oreg. Agric. Exp. Stn. 43.
Thomson J.D., Plowright R.C. (1980) Pollen carryo-ver, nectar rewards, and pollinator behavior with special reference to Diervilla lonicera, Oecologia 46, 68–74.
Torchio P.F. (1984) Field experiments with the polli-nator species, Osmia lignaria propinqua Cresson (Hymenoptera: Megachilidae) in apple orchards: III, 1977 studies, J. Kans. Entomol. Soc. 57, 517– 521.
Torchio P.F. (1985) Field experiments with the polli-nator species, Osmia lignaria propinqua Cresson in apple orchards: V, (1979–1980), methods of introducing bees, nesting success, seed counts, fruit yields (Hymenoptera: Megachilidae), J. Kans. Entomol. Soc. 58, 448–464.
Torchio P.F., Asensio E. (1985) The introduction of the European bee, Osmia cornuta Latr., into the U.S. as a potential pollinator of orchard crops, and a comparison of its manageability with Osmia
lig-naria propinqua Cresson (Hymenoptera:
Meg-achilidae), J. Kans. Entomol. Soc. 58, 42–52. Vicens N., Bosch J. (2000a) Pollinating efficacy of
Osmia cornuta and Apis mellifera (Hymenoptera:
Megachilidae, Apidae) on “Red Delicious” apple, Environ. Entomol. 29, 235–240.
Vicens N., Bosch J. (2000b) Nest site orientation and relocation of populations of the orchard pollinator
Osmia cornuta (Hymenoptera: Megachilidae),
